Influencer Finder

🔎 What can Influencer Finder do?

Keyword or niche in, creator handles out. One flat row per creator per platform found, with the handle, the profile URL, the follower count the search result showed, the search that found the creator, and up to five similar creators from the same search. Discovery only; profile detail is the Influencer Profile Scraper, and the Link in Bio Scraper and Newsletter Detector reads the link page.

Platforms: TikTok, Instagram, YouTube, Pinterest, Twitch, Threads, and podcasts (Apple and Spotify show pages, public feeds). Not X, not Facebook pages, not LinkedIn.

📦 What you get	⚙️ Features and integrations
👥 One row per creator per platform, handle and profile URL 📊 Follower count as the search result showed it 🔍 The search that found them, keyword and niche 🤝 Up to five similar creators from the same search 🧾 59 flat fields, snake_case, the suite's shared row	🌐 Seven platforms, TikTok, Instagram, YouTube, Pinterest, Twitch, Threads, podcasts 🎯 Eight curated niches, about 60 keywords each 🎚️ Follower band filter, unknown counts kept 🧪 Clay ready, one keyword per row lands on the same path ⬇️ Export to JSON, CSV, Excel, HTML or XML

Bought by teams that sell to creators, newsletter and creator tool companies, talent and brand partnership teams, and agencies building creator lists by niche.

🚫 This is not a profile scraper or a newsletter detector. It finds handles. Followers come from the search result, not the profile page. Profile detail is the Influencer Profile Scraper, and the link page is read by the Link in Bio Scraper and Newsletter Detector; the Influencer Lead List Builder runs all three in one go.

💡 Why use Influencer Finder?

If you sell	Read these fields
Newsletter or email tools to creators	handle, platform, followers, then hand the row to the Link in Bio Scraper and Newsletter Detector
Creator tools, courses, or memberships	search_keyword, niche, followers
Brand partnerships or sponsorships	followers, similar_creators, profile_url
Talent representation	handle, similar_creators, then the Influencer Talent Agency Lookup
Anything, as a disqualifier	row_status, error_reason

🧭 How discovery works, and why the yield is what it is

TikTok, Instagram, and Threads have no keyword search that renders without login, so discovery runs site:<platform> "<keyword>" "Followers" on Google through the SERP proxy and reads the handle from the result title and the follower count from the snippet. Google honors the site: operator on the first page or two and on some exit countries not at all, so the actor pages shallow, varies the exit country, and fetches a page without a handle once more before giving the keyword up. Measured on 2026-09-21: TikTok 7 handles per keyword, Instagram 1 to 8, Threads 11 to 15. YouTube (the channel filtered results page, 20 per keyword), Twitch (its own search, 11), Pinterest (its user search resource, 50), and podcasts (the iTunes Search API, 50) need no search engine.

More keywords with shallow paging beats few keywords with deep paging; that is what the niche sets are for.

similar_creators lists up to 5 other handles the same search returned (similar_creators_method is same_keyword_cluster), because every platform's related profile endpoint came back empty without login in the pre-research.

📋 What data can Influencer Finder extract?

59 fields per row, the same 59 in the same order on every actor in this suite. The ones this actor fills:

Field	What it holds
creator_id, creator_id_method, creator_id_confidence	One ID per creator across platforms and runs; see The shared row below
platform, handle, profile_url, display_name	The creator found, with the canonical public profile URL
search_keyword, niche	The search that returned this creator; niche is null for your own keywords
similar_creators, similar_creators_method	Up to 5 other handles from the same search; same_keyword_cluster
followers	The count the search result showed, rounded the way the platform rounds it
source_url, read_at, row_status, error_reason, billable_events_json	Where the row came from, when, whether it is usable, and what it cost

⚠️ false and null are not the same thing, and an empty cell is not a failed read. A column this actor does not own is null, never missing. On a boolean, false means the page was read and the signal was not there; null means nothing read it. On an Influencer Profile Scraper row, verified: false is a profile without the mark and verified: null is a profile that could not be read. Read row_status before any other column; the paragraph below says how.

Every actor in this suite writes the same 59 flat columns, in the same order, and fills the ones it owns. A column an actor does not own is null, never missing. Nested data lives only in columns ending _json, as a JSON string, so a Clay column reads one cell.

Group	Columns	Filled by
Identity	creator_id, creator_id_method, creator_id_confidence, platform, handle, profile_url, display_name	Influencer Finder, Influencer Profile Scraper
Discovery	search_keyword, niche, similar_creators, similar_creators_method	Influencer Finder; search_keyword and niche also carried over by the Influencer Profile Scraper, Link in Bio Scraper and Newsletter Detector, and Influencer Change Monitor when an input item names them
Profile	followers, following, post_count, engagement_rate, engagement_method, verified, bio, bio_link, location_text, country_guess, country_guess_method	Influencer Profile Scraper
Contact	business_email, business_email_source, website_email, website_email_source_url, manager_name, manager_email, manager_source_url, agency_name, agency_domain, agency_match_method	Influencer Profile Scraper, Link in Bio Scraper and Newsletter Detector, Influencer Talent Agency Lookup
Links	link_in_bio_platform, outbound_links_json, other_social_profiles_json, own_website, website_is_host_page	Link in Bio Scraper and Newsletter Detector; other_social_profiles_json also by the Influencer Profile Scraper
Newsletter	newsletter_status, newsletter_platform, newsletter_url, newsletter_evidence_url, newsletter_check_method	Link in Bio Scraper and Newsletter Detector
Sells	sells_course, sells_coaching, sells_digital_product, sells_merch, sells_membership, brand_deals_visible, discount_codes_visible	Link in Bio Scraper and Newsletter Detector
Change	change_type, change_from, change_to, previous_run_at	Influencer Change Monitor
Segment	segment_tier	Link in Bio Scraper and Newsletter Detector
Run	source_url, read_at, row_status, error_reason, billable_events_json	All six

The Influencer Lead List Builder runs the Influencer Finder, the Influencer Profile Scraper, and the Link in Bio Scraper and Newsletter Detector in one run and fills what they fill. The Influencer Change Monitor and the Influencer Talent Agency Lookup read the profile (and the Influencer Change Monitor reads the link page too when check_links is on), so their rows carry the identity, profile, contact, and link columns as well.

Read row_status first. ok is a normal row. error carries the reason in error_reason (here, a search that threw, with the keyword and platform on the row) and charges nothing. partial does not occur on this actor; it is the Influencer Profile Scraper's status for a profile page that was read but gave no follower count, bio, or name.

The creator ID. creator_id is cr_ plus 16 hex characters of the SHA-1 of platform:handle of the creator's primary profile, so the same creator gets the same ID in every actor and every run. When one run sees the same creator on two platforms, the rows share the ID and each row says how it was linked: bio_link_match (0.9, one profile links to the other), website_match (0.8, same own website), handle_match (0.5, same handle on two platforms, which collides on common words). A lone row is seed_handle at 1.

Country scope. Launch scope is US creators, and the filter acts at the profile read, not here. Discovery carries no country signal, so country_guess and country_guess_method are null on every row of this actor and us_only changes nothing on this actor. The field is accepted so the same input runs unchanged on the Influencer Lead List Builder, where the profile stage returns a creator whose guess is a known non US country as an error row.

🛠️ How to find influencers by keyword or niche

1. Open the Input tab and put one or more search phrases in keywords, one per line, or pick a niche to load its curated keyword set.
2. Pick platforms. The default is TikTok and Instagram; add YouTube, Pinterest, Twitch, Threads, or podcasts.
3. Set the follower band with follower_min and follower_max, or leave the defaults of 5,000 and 500,000. The band is ignored for podcasts, which carry no follower count.
4. Click Start. One row per creator per platform lands in the dataset as each search finishes.
5. Read row_status first, then handle, profile_url, and followers.
6. Export from the Output tab, or pull the rows through the API, and hand the handles to the Influencer Profile Scraper or the Link in Bio Scraper and Newsletter Detector.

🧪 Using it in Clay

Add an Enrichment > Apify column, pick this actor, and map your keyword column to keywords. Batch: one keyword is one run; a list is a batch. A Clay column sending one keyword per row lands on the same path. The 59 fields land as one flat row per creator with no reshaping, so a keyword column in Clay becomes a creator table.

📚 Batch or single

One keyword is one run; a list is a batch, and a niche loads its keyword set into the same list. Both shapes reach the same code path, and so does the shape the platform produces when Clay sends a top level array against an object schema. Keywords are deduplicated before any search, and a creator that two searches return is written and charged once. Rows are pushed as each search finishes, so a run that hits its timeout keeps every row it already resolved. A search that throws becomes an error row with the reason and the run continues.

Concurrency (batch_size) is searches in flight across all platforms, default 2, capped at 4.

Each keyword gets a time budget on each platform (keyword_time_budget_secs, default 60 seconds). A search engine page that hangs is given up at the budget, and the run starts no new keyword when its own timeout is closer than one budget, so a full 60 keyword niche set finishes as a SUCCEEDED run. The default run timeout is 7,200 seconds. The DISCOVERY_STATS record in the run's key-value store counts the searches done and lists any keyword that was not searched.

💵 How much does it cost to find influencers?

Pay per event. Every run is charged one actor-start event ($0.001) to cover the run overhead; every other event is charged for output, never for input. The run stops at the maximum cost per run you set in Apify, or at max_charge_usd, whichever comes first: it delivers every row already charged and charges nothing more.

Event	Fires when	Price
actor-start	Once per run, on start.	$0.001
creator-found	Once per creator row returned by keyword or niche discovery with a handle and a profile URL. A search that returns nothing charges nothing.	$0.007

Free Apify plan users get 200 results per calendar month, reset monthly. Upgrade to any paid Apify plan for unlimited use: https://apify.com/pricing?fpr=mamba

💳 What you are billed for. creator-found fires once per creator row with a handle and a profile URL. A search that returns nothing charges nothing, and a keyword that finds the same creator twice charges once, because input is deduplicated before any row is written. A Google page costs about $0.0025 in platform usage on your account; two pages per keyword per platform is the usual spend.

Apify bills its own platform usage on top of the event prices.

⌨️ Input

Everything is on the Input tab. The options worth explaining:

Field	What it does
keywords	Search phrases, one per line. Each is searched on each platform.
niche	Loads a curated keyword set (about 60 per niche, from the September 2026 pre-research of 1,034 creators): personal finance, business and marketing, health and fitness, beauty and skincare, parenting and family, food and cooking, travel, tech and AI. max_keywords_per_niche caps how many are used (default 10).
platforms	Which platforms to search. Default TikTok and Instagram.
max_creators_per_keyword, max_creators	Per search cap (default 20) and a hard cap on the run (default 200).
follower_min, follower_max	Drop creators whose count is known and outside the band. A result with no follower count is kept, not dropped, so a later profile read can fill it. Defaults 5,000 and 500,000. The same band is accepted under the names min_followers and max_followers as run input (API, Clay, or a task); those are not on the Input tab, and when set they win over follower_min and follower_max. The band is ignored when platforms includes podcast, for the podcast search only: podcasts carry no follower count, so every podcast row passes.
keyword_time_budget_secs	Seconds allowed per keyword per platform, default 60. Raise it to page deeper on a slow search engine; lower it to finish a large keyword set sooner.
country	Runs every search from one country with Google's matching country setting (US, GB, CA, AU, IE, NZ, DE, FR, NL, SE). Default any, which rotates the exit country.
us_only	Launch scope. Discovery carries no country signal, so this field changes nothing on this actor; it is accepted so the same input runs unchanged on the Influencer Lead List Builder, where it acts at the profile stage on country_guess.
max_charge_usd	Optional. Your own cap for this run, in USD. The run stops before any charge that would take it past this amount, delivers every row already charged, and drops the rest. Leave it empty to rely only on the maximum cost per run you set in Apify, which the actor also honors.

📤 Output

Exports to JSON, CSV, Excel, HTML or XML. One flat, snake_case row per creator per platform, 59 columns, with null rather than a missing key. No nested objects, so it drops straight into Clay, a spreadsheet or a warehouse table without a flattening step. Nested data lives only in columns ending _json, as a JSON string, so a Clay column reads one cell.

Influencer Finder fills the identity, discovery, and run columns and the follower count; the other columns are null until the Influencer Profile Scraper or the Link in Bio Scraper and Newsletter Detector fills them. The row below is trimmed to the columns this actor fills.

{
  "creator_id": "cr_4af4c45d4984f69f",
  "creator_id_method": "seed_handle",
  "creator_id_confidence": 1,
  "platform": "youtube",
  "handle": "AnudeepTraveller",
  "profile_url": "https://www.youtube.com/@AnudeepTraveller",
  "display_name": "Anudeep Budget Travel Vlogs",
  "search_keyword": "budget travel",
  "niche": null,
  "similar_creators": "Budgetfoodtravel, funandbudgettravels, travelbitsandhacks, BudgetTravels",
  "similar_creators_method": "same_keyword_cluster",
  "followers": 10300,
  "source_url": "https://www.youtube.com/@AnudeepTraveller",
  "read_at": "2026-09-22T08:19:08.409Z",
  "row_status": "ok",
  "error_reason": null,
  "billable_events_json": "[{\"event\":\"creator-found\",\"count\":1}]"
}

💡 Tips

• More keywords with shallow paging beats few keywords with deep paging; that is what the niche sets are for. Set max_keywords_per_niche up from 10 rather than max_creators_per_keyword up from 20.
• Use similar_creators as a second seed list. Each row carries up to 5 handles the same search returned.
• Keep follower_min and follower_max on. A count the search result did not show is kept, so the band never drops a creator you cannot see. On a podcast search the band does nothing, because podcasts carry no follower count.
• Run the same keywords weekly and diff handle. New handles are creators who started ranking for the phrase.

⚠️ Known limits

Discovery results vary. Influencer Finder searches public pages for each keyword, so the number of creators found per platform changes from run to run. TikTok, Instagram, and Threads discovery runs through web search and can return few or no handles on a given run. YouTube, Twitch, Pinterest, and podcasts use each platform's own search.

Discovered handles that no longer exist. Influencer Finder charges creator-found for each handle it discovers. A small share of discovered handles belong to deleted or renamed accounts. A profile read by the Influencer Profile Scraper then returns a labeled not_found row, and the discovery charge still applies.

Followers are the search result's figure. The count is read from the result snippet, so TikTok rounds above about ten thousand and YouTube rounds to three significant figures. Read the profile with the Influencer Profile Scraper for the exact count.

Discovery has no country signal. country_guess is null on every row here and us_only changes nothing on this actor; on the Influencer Lead List Builder it acts at the profile stage.

Not X, not Facebook pages, not LinkedIn. Seven platforms are searched, listed above.

What is never done. No login. No CAPTCHA solving. No LinkedIn fetch. No message to anyone. No key of ours is used on your run; this actor has no AI option and takes no API key.

What this actor shares

This run contributes the records it finds to a shared creator and agency pool that all users of this actor read from. What one run finds, the next run can read.

The toggle is contribute_to_shared_pool. It is a boolean input and it is on by default. Turn it off and the run still reads the pool and writes nothing to it.

What this actor contributes. The creator ID, platform, and handle of every creator it finds, plus the niche the run searched when one is set. No profile detail, because this actor reads none.

Only public data that is already in your own output. Every field written to the pool is a field this run returned to you, read from a page the platform or the creator publishes to anyone without a login. Nothing from your Apify account, your input list, your API keys, or your own notes is sent. A contribution never deletes anything from the pool.

What a contribution is labeled with. The actor ID, the run ID, the pool key issued to the actor build, and a hash of the calling IP address, used for the rate limit and nothing else. Your Apify account and your user ID are not recorded.

Contributing is free. No event is charged for a write to the pool. If the pool is unreachable the run finishes as normal, the rows are dropped, and the run log says so.

❓ FAQ

What is a row with no handle and run_outcome= in error_reason? It is the run's summary row. A run that delivers no creator row at all writes exactly one row that says why, so you never get an empty dataset with no explanation. row_status is partial when every source answered and nothing matched (run_outcome=no_results), and error when a platform refused the request (run_outcome=blocked) or did not answer before the run's time ran out (run_outcome=timed_out). A blocked run is never reported as zero results: error_reason names the hosts reached, the hosts that refused, the hosts that timed out, and the run ID and build. The summary row carries no billable event and is never charged.

What happens when a platform is slow or refuses the run? Every request is capped by the time left in the run, and the actor stops starting new creators 90 seconds before the run's timeout so it can write what it has. A slow or refusing platform returns that creator's labeled error row and the run carries on with the rest. The run exits successfully with the rows it found, and the status message says when it stopped early.

Why did TikTok return zero handles for my keyword? TikTok, Instagram, and Threads discovery runs through web search, and Google honors the site: operator on the first page or two and on some exit countries not at all. The actor pages shallow, varies the exit country, and retries once. A run can still return few or no handles on a given keyword; the next run, or a neighboring keyword, usually does.

Are the follower counts exact? No. They are what the search result showed. The Influencer Profile Scraper reads the profile page for the exact figure.

Does it read the profile or the link page? No. Discovery only. That is why a row costs $0.007 and not $0.021.

Why is similar_creators by shared search and not the platform's related profiles? Every platform's related profile endpoint came back empty without login in the pre-research, so similarity is by shared search (similar_creators_method is same_keyword_cluster).

Can I run a list of keywords from Clay? Yes. One keyword per row lands on the same path as a list in one run.

🧩 Want other GTM data?

Mamba Labs builds a fleet of GTM enrichment actors that share one flat, Clay-ready output convention, so their rows join on company_domain or creator_id with no cleaning step:

🕵️ Agent Accessibility Auditor	🤖 AI Tooling Detector
📡 GTM Signals Aggregator	🚀 Prospect Engine
📝 Publishing Frequency Tracker	🦋 Bluesky Brand Presence Mapper
⏩ Sequencer Lead Push	🔄 Company Change Event Feed
📇 Company Contact Details Extractor	🧭 Company Discovery List Builder
🏢 Company Firmographic Enricher	🪪 Company Identity Resolver
🌐 Company Social Presence Mapper	🏷️ Contact Classifier
📈 Influencer Change Monitor	🧾 Influencer Lead List Builder
👤 Influencer Profile Scraper	📬 Domain Deliverability Checker
🔗 Domain to LinkedIn URL Resolver	🛒 Ecommerce Platform Profiler
✉️ Work Email Waterfall	🎪 Event Presence Index
⚖️ Expert Witness Directory	💵 Company Funding Record
💰 Funding and Press Signal Scanner	🐙 GitHub Organization Signal Scanner
🧑‍💼 GTM Hiring Signal Scraper	📋 Job Posting Monitor
🧱 Tech Stack Signal Detector	🎯 ICP Fit Scorer
🔑 Job Board Keyword Signal Scanner	🏛️ Legal Entity Resolver
📰 Link in Bio Scraper and Newsletter Detector	💼 LinkedIn Company Page Mapper
💬 LinkedIn Post Tracker	📢 Meta Ad Library Monitor
📸 Meta Brand Presence Mapper	📮 Outbound Stack Detector
📄 Page Finder and Extractor	🧑 People Finder & Email Verifier
📌 Pinterest Brand Presence Mapper	🏆 Government Contract Award Monitor
📅 Reporting Window Finder	👽 Reddit Brand Presence Mapper
⭐ Trustpilot Reputation Enricher	🤝 Influencer Talent Agency Lookup
👥 Team Page People Extractor	🎵 TikTok Brand Presence Mapper
📊 Website Traffic Rank Estimator	🏅 Workplace Program Detector
✖️ X Brand Presence Mapper	▶️ YouTube Channel Stats Extractor

Every actor in the suite takes a domain, a company, or a creator and returns one flat row, so they stack in the same Clay table without reshaping anything.

🛠️ Need something custom built for you or your team? Tell us what you are trying to find and we will build it. Talk to Mamba Labs.

🆘 Support

Issues, field requests and bug reports: open an issue on the actor's Issues tab. Mamba Labs reads every one.

ℹ️ Sourcing and legal. Every field comes from pages the platforms and the creators publish to anyone without a login, read directly or through a proxy, with no login, no CAPTCHA solving, and no LinkedIn fetch. The row records what was public at read_at. Nothing is assessed or scored; a class, a flag, or a match method says what was read and where. You are responsible for how you use the output, including under applicable data protection and platform terms.

Built by Mamba Labs.